Dear STATA Intellectuals,

I am estimating the causal effect of mothers' education on children's health status using a compulsoru education law.Mothers born in 1987 and further were exposed to the law (i.e., treated) mothers born before 1986 were not exposed (not-treated). I have pooled DHS data of 2008 and 2013.

Here is my code:

ivreg2 duration311 (completionof8years=reformexposure) V012 i.child_sex age_months i.region12 i.placeofresidence V201 i.wealth_index i.birthordercat i.birthweightcat i.ceasarean [iweight=wgt], first

duration311 is an ordinal and a categorical variable indicating the durations of breastfeeding for children 0-48 months of age. It has 3 categories in terms of breastfeeding duration. (1= if child is breastfed for 0-6 months; 2= 7-12 moths; 3=13 and more months).
completionof8years is my endogenous variable which is a dummy variable whether the mothers have completed 8 years of education.
reformexposure is my instrument. It indicates whether the mother is born in 1987 and further (=1), or born before 1987 (=0). It is also a dummy variable.

When I run the abovementione code, here is what I obtained:

I can understand the first stage results. However, I do not understand where is my reduced form estimates and how do I obtain them. Can anybody please help me? What is the difference between RF and second stage?
In Wooldridge's book, I have read that first-stage is reduced form. Am I right?
I am really confused about the cases for RF and Second stage estimates.

Thank you very much in advance.

Array